Diabetes mellitus seems to be an heterogeneous disease, and associations of certain histocompatibility antigens and haplotypes with several types of diabetes or diabetic microangiopathy have been reported. We have hypothesized that there is one or more "diabetic genes" in linkage disequilibrium with histocompatibility (HLA) genes with important roles in the pathogenesis of juvenile, insulin-dependent diabetes, maturity-onset type of diabetes of young people, and in severe diabetic microangiopathy. We further hypothesize that certain HLA haplotypes predict the development of diabetes. To test these hypotheses we plan to HLA genotype for loci A, B, C, and D 120 families with juvenile, insulin-dependent diabetes, 10 families with conjugal diabetes, and 20 families with maturity-onset type of diabetes of young people and to study linkage analysis between hypothetical "diabetic genes" and the HLA genes. Adequate samples of unaffected sibs of the diabetics, both HLA identical and HLA nonidentical with their diabetic siblings, will be followed prospectively for five years to test the potency of HLA haplotypes as predictors of diabetes. To study the associations between HLA antigens and diabetic microangiopathy, we will tissue type several samples (about 100 subjects each) of unrelated juvenile, insulin-dependent, and adult-onset, insulin-independent diabetics with either severe microangiopathy. We also plan to study B cell alloantigens in both the family and population studies, as a guest for further indicators useful in pathogenetic studies of diabetes.